Visual examination of the fuel assembly is used to:

• Get a general impression of the condition of the fuel assembly.

• Assess zirconium oxide type and whether oxide spallation has occurred or not.

• Identify the primary failure cause and characterize degradation of failed fuel.

• Measure component dimensions by means of translation stages and position encoders.

Fuel assemblies are sometimes inspected while supported on the fuel han­dling mast. More commonly, visual inspections are performed in fuel prepa­ration machines located in fuel storage pools. The fuel preparation machines are usually equipped with a rotating fixture to enable inspections of all four assembly faces without the use of the bridge crane and with a sensor to rec­ord axial position.

Visual inspection equipment, typically consists of:

• Television camera and spotlights.

• In-pool support system for camera and lights.

• Data acquisition, processing and storage equipment.

The second law of thermodynamics is an expression of the tendency over time differences in temperature, pressure and chemical potential will equili­brate in an isolated physical system. In other words, every material tends to reach the maximum of disorder, in order to minimize its potential energy. With regard to corrosion, it means that leaving the crystalline network under the action of an electric field, metal ions yields energy. According to thermodynamics, almost all metals have negative free energy, suggesting their reactivity in environments where they are exposed.

Corrosion reactions are electrochemical in nature, based on mass and charge transfers. Reactions can be split into partial oxidation and reduction reactions. The potential is the propensity to exchange electrons: the metal donating electrons is oxidized, while the metal receiving electrons is reduced. The stability of elements in a given medium is predicted by the correspond­ing Pourbaix diagram, where predominant phases are defined in agreement with thermodynamics. However, reaction kinetics play a major role in the evolution of the system (such as changes in pH, potential or temperature).

For example, if iron is introduced into hydrogenated water at 300°C (with­out any dissolved oxygen) at pH 7 and a potential of -700 mVSHE, cations Fe2+ are dissolved in the water (Reaction [2.1]). This anodic reaction (oxi­dation of metal) is coupled to the cathodic reaction (reduction of water) described by Reaction [2.2]. Then, dissolved cations Fe2 + can join oxidant ions OH — (Reaction [2.3]), to form ferrous hydroxide Fe(OH)2.

Cations can be released in the water due to dissolution. When the satura­tion in ferrous hydroxide Fe(OH)2 or ferrous cation Fe2+ is reached, accord­ing to the Pourbaix diagram, magnetite Fe3 O4 forms based on Schikorr Reaction [2.4].

3Fe(OH)2 -» Fe3O4 + H2 + 2H2O [2.4]

Finally, passivation is the process of building a protective layer of oxide isolating the surface of the material from the aggressive environment. Some corrosion inhibitors help the formation of such layers. Figure 2.1 shows energy-dispersive x-ray (EDX) analysis of the oxide formed on stainless steel exposed to water at 360°C (-600 mVSHE, pH325°C = 7.2). The passive film is the 50 nm-thick layer containing a significant level of chromium at the surface of the metal.

In Section 3.1 we identified stress, temperature and grain size/microstructure as the three important factors which determine the extent of creep defor­mation that a material experiences. However there are examples where two materials with similar compositions, grain sizes and second phase distributions might creep at vastly different rates under a given stress and temperature. Such anomalous behavior has been observed in titanium alloys by Mishra et al." who found that alloys with nominally similar compositions crept at sig­nificantly different rates. It was found that the presence of trace elements such as Fe and Ni degrade the creep properties of the titanium alloy. Even though these elements are present only in the order of ppm, they influence the dif­fusion rates to an extent as to bring about significant changes in creep rates. The activation energy for diffusion in the higher Fe/Ni containing alloys was found to be smaller and vice versa. Mishra et al. suggest that the Fe/Ni appear to dissolve interstitially and form foreign atom-vacancy pairs which play a sig­nificant role in accelerating the diffusion kinetics of the titanium.

Irradiation growth of RXA Zr-Nb alloys (E110, E635, NSF, M5, ZIRconium Low Oxidation (ZIRLO)) all exhibit a resistance to formation of <c> loops at low or intermediate fluence and as a result have lower growth than

Zircaloys. The fluence to breakaway is not yet well defined but is probably at least 1.5 x 1022 n/cm2, E>1 MeV (see Fig. 4.63) corresponding to a burnup of about 75 MWd/kgU.

The IAEA has stated that electrical properties — voltage withstand, insula­tion resistance, capacitance, attenuation, and/or signal propagation — are the ‘most important functional properties’ of cables (IAEA, 2011). Such prop­erties provide a direct measure of the loss of cable resistance or dielectric parameters and therefore its loss of functionality (IAEA, 2011). Electrical measurements are primarily suitable for cable conductors, connectors, splices and penetrations. Most electrical measurement techniques are less sensitive to problems with cable insulation material though they can reveal them (IAEA, 2011).

Two types of electrical tests are available: destructive methods, which iden­tify cable failure locations before the cable is installed, and non-destructive methods, which are better suited to identifying cable degradation (U. S. NRC, 2001). Non-destructive tests can be categorized by direct current (DC) and alternating current (AC) methods. DC tests generally require the least expen­sive test equipment, but may be less appropriate for some power cables and cables used in AC applications (U. S. NRC, 2001). The primary advantage of electrical techniques is that they can be used in-situ on installed and less accessible cables, providing information on the entire length of a cable, not just those points which are tested (IAEA, 2011). Because some methods also enable trending based on baseline measurements, electrical techniques can be used to note changes over time for ageing management purposes (IAEA, 2011). The most important electrical parameters in cables are insu­lation resistance, leakage current, loss factor, permittivity, and breakdown voltage. Provided that one or both ends of the cable are accessible to mea­surement and the cable can be de-energized, these electrical parameters can usually be measured on any cable (Hashemian, 2010).

296 Materials’ ageing and degradation in light water reactors There are two general types of in-situ electrical cable tests:

1. Insulation quality tests, which include insulation resistance (IR), high-potential (Hi-Pot), partial discharge, quality factor, dissipation fac­tor, and AgeAlertTM.

2. Impedance tests, which include LCR (inductance, capacitance, and resis­tance), time domain reflectometry (TDR), and frequency domain reflec — tometry (FDR).

Many of these electrical tests are simple and have been in use for decades. In recent years, LCR measurements have been added to the TDR test to improve cable diagnostics, help identify the nature of a fault, and pinpoint its location along a cable. Used together, electrical methods like the TDR and LCR tests provide an overall picture of cable health as well as informa­tion for expediting any repairs that may be needed (Hashemian, 2010).

Many conventional electrical test methods such as IR and LCR are only used to give a snapshot of the current condition of a cable. Others, such as TDR, FDR and RTDR, can identify the fault location within the length of cable, but may not differentiate whether the problems are in the connection or the end device. Additional tests are normally required to help distinguish whether the fault is in the cable or to diagnose the cause of the end device problem (AMS Corp., 2011). In recent years, the TDR and FDR techniques have been either combined or packaged and introduced under such names as LIRA (line impedance resonance analysis) a method which seems to be essen­tially the same as the FDR technique and JTFDR (joint time and frequency domain reflectometry) which combines TDR and FDR in a single test.

The analysis of any creep data is made by assuming the microstructure to be constant. Some of the microstructural features that could change during the course of a test are phase composition, precipitate size and distribution, and grain size. Thus to estimate the different creep parameters and to deter­mine the mechanism of creep, it is necessary to keep the microstructure constant. To this end, materials are usually heat treated at temperatures higher than the test temperature. Even though thermal stabilization estab­lishes a constant microstructure during the course of a test, stress-assisted processes altering the microstructure cannot be ruled out. Non-equilibrium structures, namely nanocrystalline materials undergo stress-assisted micro­structural changes that prevent the attainment of a constant creep micro — structure.14 Creep tests on such materials should be carried out at stresses lower than the critical stress at which microstructural changes could be initiated.15

The most important microstructural parameter that plays a major role in controlling the creep properties of a material is the grain size. The depen­dence of the strength of a material on its grain size can be understood through the Hall-Petch relationship which states that materials with finer grain sizes possess greater strength than materials with larger grain size3:

[3.15]

where ay is the yield strength, a0 the friction hardening (stress felt by dislo­cations while moving through the lattice), d the grain size and ky is known as the Petch-unpinning coefficient that frees the dislocations locked by inter­stitial solute atoms. This is true at low temperatures where grain boundary sliding is not dominant.

On the contrary, under creep conditions the reverse is true. Materials with finer grain size creep faster than coarse grained materials at higher tem­peratures and at lower stresses. There are certain creep mechanisms that operate faster in finer grained materials in comparison to coarse grained materials. Hence, it is necessary to have knowledge of the grain size of a material. The dependence of the steady-state creep rate on the grain size is understood through the following equation:

[3.16 ]

where K2 is a constant, d is the grain size and p is the grain size exponent. As is clear from this equation at a given stress and temperature, finer grain sized materials are expected to creep faster than coarser grained materials. However for dislocation-based mechanisms which are not grain size depen­dent, the strain rate of deformation would be the same for both fine grained and coarse grained materials.

There are essentially, four different types of commercial light water cooled reac­tors, whose main characteristics are provided in Table 4.1 (Cox et al, 2006).

4.1.1 Pressurized water reactors (PWRs) and boiling water reactors (BWRs)

There is a wide variety of fuel assembly (FA) types for BWRs and PWRs. The fuel rod array for BWRs was initially 7 x 7 but there has been a trend over the years to increase the number of FA rods and today most designs

* Variation from lower to upper part of the core and from plant to plant. Source: A. N.T International (2011) and Cox et al. (2006).

are either of 9 x 9 or 10 x 10 square configuration design (Cox et al., 2006). The driving force for this trend was to reduce the linear heat generation rate (LHGR), which resulted in a number of fuel performance benefits such as lower fission gas release (FGR) and increased pellet clad interaction (PCI) margins. However, to increase utility competitiveness, the LHGRs of 9 x 9 and 10 x 10 FA have successively been increased, and peak LHGRs are today almost comparable to those of the older 7 x 7 and 8 x 8 designs.

Also for PWRs there has been a trend to greater subdivision of fuel rods, for example from the Westinghouse 15 x 15 to 17 x 17 design (Cox et al, 2006). However, since PWRs do not have the same flexibility with core internals and control rods as BWRs, to accomplish this requires modifica­tion of the reactor internals. There is, however, one exception, namely DC

Cook 1, which is switching to 17 x 17 through changing the reactor internals. Figure 4.2 shows the current PWR fuel rod array designs.

In most PWRs, the assemblies are positioned in the core by bottom and top fittings, and the lateral clearances are restricted by the assembly-to-assembly contacts at the spacer-grid levels (Cox et al, 2006). Furthermore, the control rods consist of rod cluster control assemblies (RCCAs), the poison part of which moves into guide thimbles (GTs). These guide thimbles are an inte­gral part of the assembly structure.

In all BWRs, the assemblies are enclosed in ‘fuel channels’ surrounding the assemblies and between which the blades of the control rods moves.

Irrespective of the many possible different shapes, sizes and configura­tions, the common FA design requirements are (Cox et al, 2006):

• Maintain proper positioning of the fuel rods under normal operating conditions and in design basis accidents (DBAs) (e. g. seismic effects, LOCA, RIA).

• Permit handling capability before and after irradiation.

Figures 4.3 and 4.4 show a typical BWR and PWR FA, respectively. Also, the different FA components are shown and the material selections for these

Rod cluster control

Top nozzle

304 L stainless steel

springs

inconel 718

Control rod

304 L stainless steel

clad

Bottom view

4.4 Typical PWR FA (Cox et al., 2006).

components are provided. The selection of the different structural materi­als is based on their nuclear and mechanical properties as well as their cost, in order to ensure acceptable performance during normal operation and accidents.

4.5

Draft of the RBMK-1500 fuel assembly. (Cox et al., 2006).

The LOCA event starts with a decrease and then the loss of coolant flow due to a break in the coolant pipe; at the same time the reactor is depres­surized, scrammed and shut down (Strasser et al, 2010b). The fuel starts heating up due to its decay heat until the emergency core cooling systems (ECCSs) are activated and fuel cooling commences. Hypothetical LOCA events are analyzed for each reactor to ensure that the safety criteria, as defined by the regulators for the reactor system and the fuel, are met. The design basis accidents (DBAs) which are analyzed fall into two gen­eral categories. The large break, or large break loss of coolant accident (LBLOCA), assumes a double ended break of a primary coolant cold leg of a PWR or a break in the recirculation pump intake line of a BWR, either of which could cause the loss of all the coolant from the core. The small break, or small break LOCA (SBLOCA), assumes a break in one of the smaller primary circuit lines that will cause less coolant loss than the LBLOCA.

The effect of a LOCA cycle on the fuel is shown schematically in Fig. 5.9, plotting the fuel and cladding temperatures as a function of time in the acci­dent (Strasser et al. , 2010b). The loss of coolant flow and reactor pressure at the initiation of the accident will decrease heat transfer and allow the fuel and cladding to heat up until the reactor scrams. The fuel will then cool

о
400
Coolant blockage
50	100	150

Time (s)

5.9 Typical LOCA in a PWR (Strasser et al., 2010b).

down somewhat, partly due to cooling by the steam-water mixture that is

formed, but the cladding temperature will continue to rise.

During and after the LOCA it must be ensured that (Strasser et al,

2010b):

• The core remains coolable (which means that the maximum allowable coolant blockage is limited)

• No fuel dispersal occurs (which means that cladding rupture is not allowed; it is assumed that the cladding burst is so small that only fission gases are released)

• Less than 10% of the fuel rods in the core fail through burst (but without fuel dispersal) (a requirement in Germany only).

The n = 3 regime, though in principle corresponding to the power-law con­trolled (n = 4-1) creep mechanism, differs from it at a mechanistic level. The power law controlled creep mechanism (as will be discussed in the fol­lowing section) is mostly dislocation climb-controlled commonly noted in pure metals and class-II or metal-class alloys. In contrast the n = 3 regime is dislocation glide-controlled creep usually exhibited by alloys known as

class-I or class-A, and hence the n = 3 regime at times is referred to as alloy type creep behavior. The creep behavior of materials can thus be classified into two groups: class-A and class-M. The creep behavior of solid solutions or class-A alloys at intermediate stresses and for specific material param­eters consists of 3 regimes. As shown in Fig. 3.10 , as stress increases the
stress exponent changes from 5 (region I) to 3 (region II) and back to 5 (region III).

The creep behavior illustrated in Fig. 3.10 is a consequence of a compe­tition between two rate controlling mechanisms: dislocation climb and dis­location glide. Once the dislocations are generated from Frank-Read (FR) sources on parallel glide planes, the leading edge dislocations first glide and then climb to annihilation. In pure metals dislocation glide is relatively faster compared to the diffusion-controlled climb and thus climb becomes the rate controlling process resulting in n = 5. In class-A alloys, the rate of glide is controlled by the diffusion of the solute atoms, thereby leading to a relatively slower rate of glide compared to that of climb whereby the vis­cous glide of dislocations becomes the rate controlling process with n = 3; this mechanism is known as Weertman microcreep.54 Region II, the three power-law creep regime, is also known as the viscous glide regime. Viscous glide is described by

. 0.35 n a

єs = Ds I

s A s I E

where A is an interaction parameter that depends upon the viscous process controlling dislocation glide and Ds is the solute diffusivity.

The viscous process can be of different types. According to Cottrell and Jaswon,58 the dragging force could be due to the segregation of solute atmo­spheres to moving dislocations. The dislocation speed in this case is con­trolled by the rate of migration of the solute atoms. Fisher55 suggested that the viscous process had its origin in the destruction of the short range order in solid solution alloys. The disorder created by dislocation motion would result in the formation of a new interface thereby the interfacial energy becomes the rate controlling process. Suzuki59 suggested that the drag­ging force was an outcome of solute atoms segregating to stacking faults. There are suggestions that the obstacle to dislocation motion could be the stress-induced local ordering of solute atoms. The ordering of the region surrounding a dislocation reduces the total energy of the crystal pinning the dislocation.

The three power-law creep region has usually been observed to occur in solid solutions with a large atom size mismatch. Alloys with higher con­centrations of the solute atoms seem to prefer the three power-law creep regime as a viable creep mechanism. In fact, for very high concentrations of the solute atoms, regime II could be suppressed. In addition, class-A alloys usually exhibit either no or little primary creep or a region characterized by an increasing slope (increasing strain rate). This is in sharp contrast to pure metals and class-M alloys that exhibit a distinct primary creep curve with a decreasing strain rate; distinguishing features of class-A and class-M alloys

2.11 Deformation microstructure in Nb-modified Zr-alloy crept in the three power-law regime.61

were summarized by Murty.60 Some of the different alloys that exhibit three power-law creep behavior are Al-Zn, Al-Ag, and Ni-Fe alloys.

Microstructural features

Since the n = 3 region is dislocation glide-controlled, recovery based pro­cesses (such as climb) are considered to be less important. The deformation microstructures are found to consist of a large number of dislocations as shown in Fig. 3.11. In comparison to class-M alloys, the deformation micro­structures of class-A alloys are devoid of subgrains.61

A brief summary of hydride effects is given here to provide background for pellet-cladding mechanical interaction (PCMI) type failures. All zirconium alloy reactor components absorb hydrogen during reactor service through the corrosion reaction between zirconium and water. Basics of these phenomena are given in ZIRAT Special Topical Reports (Cox & Rudling, 2000; Adamson et al, 2006; Strasser et al, 2008). Hydrides tend to embrittle zirconium alloys and therefore their effects are important for in-reactor normal service, for ex-reactor handling operations and for accident and transient scenarios such as LOCA and RIA. It is thought that individual hydrides themselves are actu­ally brittle at all normal reactor temperatures (Simpson & Cann, 1979; Shi & Puls, 1999); and it is clear that high concentrations of hydrides (5000-16000 ppm) are very brittle, as in hydride blisters or rims.

Under normal conditions, hydride platelets form in the circumferential direction in fuel cladding illustrated in Fig. 4.27a, but under some circum­stances such as during long term storage or during power transients they

4.27 Hydride orientation in Zircaloy-4 (SRA) cladding: (a) circumferential, (b) radial (Chu et al., 2005).

can form in the radial direction (Fig. 4.27b). Because in high power rods a temperature gradient encourages hydrogen to diffuse to the colder outer clad surface, rims of hydrides can form, illustrated in Fig. 4.28a.

Hydrides effects are listed here, giving appropriate figures and references.

• The effect of hydrides is strongly dependent on testing temperature. Material at 300°C (573K) (reactor operating temperature regime) retains much more ductility than at 20°C. Figures 4.28b and 4.29 indicate the ductile-to-brittle transition for unirradiated material is less than 200°C, for circumferentially oriented hydrides. Figures 4.30 and 4.31 indicate that at 332°C the primary reduction in ductility comes from the irradia­tion effect, while at room temperature the effect on ductility of irradia­tion and hydrides is additive for uniformly distributed hydrides below about 1000 ppm. It is apparent that below 100°C ductility is very low.

• The distribution of hydrides is important. Dense layers of hydrides (for instance at fuel cladding surfaces) retain little ductility at any tempera­ture, and are susceptible to crack formation. Whether or not the crack will be arrested by the relatively ductile zirconium matrix depends on the layer thickness, as shown in Fig. 4.32.

• The strength of irradiated or unirradiated Zircaloy is insensitive to hydrogen content. See Fig. 4.33.

• Existence of radial hydrides can substantially reduce ductility, particu­larly at room temperature. Figure 4.34 shows the failure strains for the range of hydride orientations given in Fig. 4.35. When radial hydrides exist as in Fig. 4.35c failure strain is low. Figure 4.36 indicates that a high percentage of radial hydrides reduces the failure strain at room

As-received 200 ppm 400 ppm 500-650 ppm

800-950 ppm 1000-1300 ppm 1300-1450 ppm >1550 ppm

temperature but not at 300°C (573K). All specimens are unirradiated and are tested with applied stress normal to the hydride platelet. For similar materials having the applied stress parallel to the hydride plate­let, no hydride effect is seen (Yagnik et al, 2004).